Collective Error = Average Individual Error – Prediction Diversity

No Gravatar

The Brain has a comment on Scott Page&#8217-s presentation [PPT file] about his book, The Difference:

One question is whether there is a way to usefully systematize that principle by trader selection or via something other than a strict market such as an (incentivised) weighted average system like HP’s BRAIN. The latter type of system will work better in some domains but in general might be less robust because it constantly runs the danger of being overfit to past trader correlations (or apparent expertise, risk appetites, etc). In general, trader self-selection and self-weighting might be more accurate. If you have some metric of “fundamental” trader similarity like proximity, org chart relations, demographic data – as opposed to past trading correlations only, that might work better in terms of expert-selection/expert-weighting.

The Difference

Previous blog posts by Chris F. Masse:

  • BetFair-TradeFair has won its second Queen’s Award for Enterprise in its eight-year history.
  • Inkling Markets is one of the “Hot Companies To Watch In 2008”, according to Forrester.
  • Plenty of great news coming from Inkling Markets in the coming weeks
  • ??? charity-driven prediction markets OR social issue prediction markets ???
  • That can’t be Nigel Eccles of HubDub.
  • The Marketing Of The Reading Of The Public Prediction Markets = What Robin Hanson has deep trouble with, and what the prediction exchanges (e.g., InTrade-TradeSports, BetFair-TradeFair) haven’t fully computed yet
  • In 2013, Enterprise 2.0 will be a $4.6 billion industry. Good. But they forgot to mind the enterprise prediction markets.

Leave a Reply

Your email address will not be published. Required fields are marked *